Image Processing Apparatus for Automatic Water Drop Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing camera systems on vehicles face challenges in accurately recognizing road markings due to water droplets on the lens, leading to decreased accuracy in vehicle positioning, as manual removal methods are inefficient and require operator intervention.
Innovation Solution
An image processing apparatus that detects candidate regions for water droplets based on edge strength in captured images, estimates circles to determine if pixels within those regions have low edge strength, and automatically determines if a water droplet is present on the lens, triggering an automated removal system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual operation is used to remove water drops by outputting compressed air, then water drops can be removed from the lens, but it is very difficult to keep the lens free from water drops continuously
Solution Approach 1:
The system automatically detects water drops on the lens using image processing and triggers compressed air output without manual intervention. The microcomputer continuously monitors captured images, identifies water drop regions based on edge strength analysis, and activates the water drop removing apparatus autonomously, enabling the system to serve itself
Solution Approach 2:
The system uses real-time image capture and analysis to detect water drops, then feeds this information back to control the compressed air output. The microcomputer processes captured images, determines water drop presence through edge strength evaluation, and adjusts the water drop removing apparatus operation accordingly, creating a closed-loop feedback system
2Productivity
If manual operation is used to remove water drops, then water drops can be removed, but operator intervention is required which reduces efficiency
Solution Approach 1:
The system performs automatic water drop detection and removal without requiring operator intervention. The microcomputer autonomously processes captured images, identifies water drops through edge strength analysis, and controls the compressed air output mechanism, eliminating the need for manual operation while maintaining removal effectiveness
Solution Approach 2:
The system replaces manual mechanical operation with an automated image processing and control system. Instead of requiring operators to manually operate the compressed air device, the patent uses a microcomputer-based system that captures images, processes them through edge detection algorithms, and automatically controls the water drop removal mechanism
3Duration of action of stationary object
If water drops are present on the lens, then image capture continues, but accuracy of recognizing vehicle position decreases
Solution Approach 1:
The system takes preliminary action to remove water drops before they significantly degrade image quality and affect vehicle position recognition. By continuously monitoring captured images for water drops and automatically triggering removal operations, the system prevents water drops from accumulating to levels that would compromise measurement precision
Solution Approach 2:
The system maintains continuous image capture capability while simultaneously performing automatic water drop detection and removal. The microcomputer processes images in real-time, identifies water drops, and activates the removal apparatus as needed, ensuring that image capture operations continue uninterrupted with maintained accuracy
Data Source
AI summary
An image processing apparatus includes a detector, an estimator and a determiner. The detector detects a candidate region of a captured image captured by a camera, the candidate region serving as a candidate for a water drop region affected by a water drop on the lens of the camera, based on an edge strength of each pixel in the captured image. The estimator estimates, based on the candidate region, a circle that includes the candidate region. The determiner determines whether or not the candidate region is part of the water drop region based on the edge strength of some of the pixels in the circle.